I have realized a system which allows the retrieve of the similar images given a query image. I have implemented 4 methods of features extraction:

• SIFT with BOVW
• Daisy
• Deep Method (CNN)
• Color based

Given the query image:

1. I classify it with trained and fine-tuned Convolutional Neural Network
2. I extract, with these 4 methods, the features of each image contained into the dataset
3. I compute the similarity between the features of the query image and those of the dataset

The features of each image contained into the dataset are previously computed and then saved.